INDEX

Explanations

references to news agencies and media sources, particularly those related to Iran and Russia

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

egral

-0.16

adro

-0.16

Codec

-0.15

 Marino

-0.15

aldi

-0.14

 ActionTypes

-0.14

Ø§Ú©Ùħ

-0.14

ardown

-0.14

ernals

-0.14

Wiki

-0.14

POSITIVE LOGITS

 agency

0.37

agency

0.33

 Agency

0.33

Agency

0.30

 wire

0.29

 bureau

0.26

 agencies

0.26

 Agencies

0.23

 wires

0.23

GENCY

0.21

Activations Density 0.055%

references to news agencies and media sources, particularly those related to Iran and Russia

No Comments

No Known Activations